COMPARISON OF METHANE AND NATURAL GAS COMBUSTION BEHAVIOR AT GAS TURBINE CONDITIONS WITH FLUE GAS RECIRCULATION

The efficiency and economics of carbon dioxide capture in gas turbine combined cycle power plants can be significantly improved by introducing Flue Gas Recirculation (FGR) to increase the CO(2) concentration in the flue gas and reduce the volume of the flue gas treated in the CO(2) capture plant [1], [2]. The maximum possible level of FGR is limited to that corresponding to stoichiometric conditions in the combustor. Reduced excess oxygen, however, leads to negative effects on overall fuel reactivity and thus increased CO emissions. Combustion tests have been carried out in a generic burner under typical gas turbine conditions with methane, synthetic natural gas (mixtures of methane and ethane) and natural gas from the Swiss net to investigate the effect of different C(2+) contents in the fuel on CO burnout. To locate the flame front and to measure emissions for different residence times a traversable gas probe was designed and employed. Increasing the FGR ratio led to lower reactivity indicated by a movement of the flame front downstream. Thus, sufficient flame burnout indicated by low emissions of unburned components (CO, UHC)-required a longer residence time in the combustion chamber. Adding C(2+) or H(2) to the fuel moved the flame zone back upstream and reduced the burnout time. Tests were performed for the various fuel compositions at different FGR ratios and oxidant preheat temperatures. For all conditions the addition of ethane (6 and 16 % vol.) or hydrogen (20 % vol.) to methane shows comparable trends. Addition of hydrogen to (synthetic) natural gas which already contains C(2+) has less of a beneficial effect on reactivity and CO burnout than the addition of hydrogen to pure methane. A simple ideal reactor network based on plug flow reactors with internal hot gas recirculation was used to model combustion in the generic combustor. The purpose of such a simple model is to generate a design basis for future tests with varying operating conditions. The model was able to reproduce the trends found in the experimental investigation, for example the level of H(2) required to offset the effect of oxygen depletion due to simulated FGR.